Papers by Abdulrahman Ibraheem
arXiv (Cornell University), Dec 21, 2014
The feature extraction problem occupies a central position in pattern recognition and machine lea... more The feature extraction problem occupies a central position in pattern recognition and machine learning. In this concept paper, drawing on ideas from optimisation theory, artificial neural networks (ANN), graph embeddings and sparse representations, I develop a novel technique, termed SENNS (Sparse Extraction Neural NetworkS), aimed at addressing the feature extraction problem. The proposed method uses (preferably deep) ANNs for projecting input attribute vectors to an output space wherein pairwise distances are maximized for vectors belonging to different classes, but minimized for those belonging to the same class, while simultaneously enforcing sparsity on the ANN outputs. The vectors that result from the projection can then be used as features in any classifier of choice. Mathematically, I formulate the proposed method as the minimisation of an objective function which can be interpreted, in the ANN output space, as a negative factor of the sum of the squares of the pair-wise distances between output vectors belonging to different classes, added to a positive factor of the sum of squares of the pair-wise distances between output vectors belonging to the same classes, plus sparsity and weight decay terms. To derive an algorithm for minimizing the objective function via gradient descent, I use the multi-variate version of the chain rule to obtain the partial derivatives of the function with respect to ANN weights and biases, and find that each of the required partial derivatives can be expressed as a sum of six terms. As it turns out, four of those six terms can be computed using the standard back propagation algorithm; the fifth can be computed via a slight modification of the standard backpropagation algorithm; while the sixth one can be computed via simple arithmetic. Finally, I propose experiments on the ARABASE Arabic corpora of digits and letters, the CMU PIE database of faces, the MNIST digits database, and other standard machine learning databases.
arXiv (Cornell University), Jun 9, 2015
arXiv (Cornell University), Apr 11, 2022
arXiv (Cornell University), Dec 21, 2014
Neural Processing Letters, Jun 1, 2019
Recent systems from premier research labs, such as Facebook's and Google's, employ variants of th... more Recent systems from premier research labs, such as Facebook's and Google's, employ variants of the basic siamese neural networks (SNNs), a testimony to how SNNs are becoming very important in practical applications. The objective function of an SNN comprises two terms. Whereas there are no issues about the choice of the first term, there appears to be some issues concerning the choice of the second term, along the lines of: 1. apriori boundedness from below; and 2. vanishing gradients. Therefore, in this work, I study four possible candidates for the second term, in order to investigate the roles of apriori boundedness from below, and vanising gradients, on classification accuracy, as well as to, more importantly, from a practical standpoint, elucidate the effects, on classification accuracy, of using different types of second terms in SNNs. My results suggest that neither apriori boundedness nor vanishing gradients are crisp decisive factors governing the performances of the candidate functions. However, results show that, of the four candidates evaluated, a particular candidate features generally superior performance. I therefore recommend this candidate to the community, and this recommendation attains especial importance when taken against a backdrop of another facet of this work's results which indicates that choosing a wrong objective function could cause classification accuracy to dip by as much as 17%.
ArXiv, 2015
We propose a novel technique, termed compact shape trees, for computing correspondences of single... more We propose a novel technique, termed compact shape trees, for computing correspondences of single-boundary 2-D shapes in O(n2) time. Together with zero or more features defined at each of n sample points on the shape's boundary, the compact shape tree of a shape comprises the O(n) collection of vectors emanating from any of the sample points on the shape's boundary to the rest of the sample points on the boundary. As it turns out, compact shape trees have a number of elegant properties both in the spatial and frequency domains. In particular, via a simple vector-algebraic argument, we show that the O(n) collection of vectors in a compact shape tree possesses at least the same discriminatory power as the O(n2) collection of lines emanating from each sample point to every other sample point on a shape's boundary. In addition, we describe neat approaches for achieving scale and rotation invariance with compact shape trees in the spatial domain; by viewing compact shape tree...
Towards Reliable Handwritten Character Recognition amid Diacritic Chaos
The orthographies of several languages, such as Yoruba, Arabic, Tshivenda, Ciluba, French, German... more The orthographies of several languages, such as Yoruba, Arabic, Tshivenda, Ciluba, French, German, and Dutch, use diacritics. These diacritics pose an additional challenge to a character recognition system. Using diacritically-marked uppercase Yoruba letters as a case-study, this paper presents one strategy for addressing this problem. In particular, we present a system for the automatic classification of diacritically-marked handwritten uppercase Yorùbá letters in offline mode. Our approach involves six stages: a pre-processing stage; a segmentation stage for isolating the pertinent Latin letter from the diacritical mark(s); a feature extraction stage where eight geometric properties of the Latin letter are computed; a Bayesian classifier stage where the Latin letter is classified based on the extracted features; a decision tree stage where the diacritical marks are recognized; and a result fusion stage where the results of the two latter stages are combined into a single final cla...
A maiden edition of AUSTECH 2015 International Conference Book of Abstracts
This maiden edition of the “Book of Abstracts” contains the abstracts of the various presentation... more This maiden edition of the “Book of Abstracts” contains the abstracts of the various presentations in the 2015 AUSTECH International conference by distinguished professors in the fields of Computer Science, Materials Science & Engineering and Petroleum Engineering. Each field as a track has different activities which includes poster presentations, PhD paper contests, technical sessions and panel discussions. These presentations were done by distinguished experts in the field from within Nigeria and across the globe. AUSTECH is an annual event by AUST. It focuses on current developments in Engineering technologies, scientific and industrial applications for development in Sub-Saharan Africa.
1 A Concept Paper on a New Model for Automatic Diacritic Restoration
Neural Processing Letters, 2018
Recent systems from premier research labs, such as Facebook's and Google's, employ variants of th... more Recent systems from premier research labs, such as Facebook's and Google's, employ variants of the basic siamese neural networks (SNNs), a testimony to how SNNs are becoming very important in practical applications. The objective function of an SNN comprises two terms. Whereas there are no issues about the choice of the first term, there appears to be some issues concerning the choice of the second term, along the lines of: 1. apriori boundedness from below; and 2. vanishing gradients. Therefore, in this work, I study four possible candidates for the second term, in order to investigate the roles of apriori boundedness from below, and vanising gradients, on classification accuracy, as well as to, more importantly, from a practical standpoint, elucidate the effects, on classification accuracy, of using different types of second terms in SNNs. My results suggest that neither apriori boundedness nor vanishing gradients are crisp decisive factors governing the performances of the candidate functions. However, results show that, of the four candidates evaluated, a particular candidate features generally superior performance. I therefore recommend this candidate to the community, and this recommendation attains especial importance when taken against a backdrop of another facet of this work's results which indicates that choosing a wrong objective function could cause classification accuracy to dip by as much as 17%.
The feature extraction problem occupies a central position in pattern recognition and machine lea... more The feature extraction problem occupies a central position in pattern recognition and machine learning. In this concept paper, drawing on ideas from optimisation theory, artificial neural networks (ANN), graph embeddings and sparse representations, I develop a novel technique, termed SENNS (Sparse Extraction Neural NetworkS), aimed at addressing the feature extraction problem. The proposed method uses (preferably deep) ANNs for projecting input attribute vectors to an output space wherein pairwise distances are maximized for vectors belonging to different classes, but minimized for those belonging to the same class, while simultaneously enforcing sparsity on the ANN outputs. The vectors that result from the projection can then be used as features in any classifier of choice. Mathematically, I formulate the proposed method as the minimisation of an objective function which can be interpreted, in the ANN output space, as a negative factor of the sum of the squares of the pair-wise distances between output vectors belonging to different classes, added to a positive factor of the sum of squares of the pair-wise distances between output vectors belonging to the same classes, plus sparsity and weight decay terms. To derive an algorithm for minimizing the objective function via gradient descent, I use the multi-variate version of the chain rule to obtain the partial derivatives of the function with respect to ANN weights and biases, and find that each of the required partial derivatives can be expressed as a sum of six terms. As it turns out, four of those six terms can be computed using the standard back propagation algorithm; the fifth can be computed via a slight modification of the standard backpropagation algorithm; while the sixth one can be computed via simple arithmetic. Finally, I propose experiments on the ARABASE Arabic corpora of digits and letters, the CMU PIE database of faces, the MNIST digits database, and other standard machine learning databases.
arXiv (Cornell University), Apr 11, 2022
Subtle and overt racism is still present both in physical and online communities today and has im... more Subtle and overt racism is still present both in physical and online communities today and has impacted many lives in different segments of the society. In this short piece of work, we present how we're tackling this societal issue with Natural Language Processing. We are releasing BiasCorp, a dataset containing 139,090 comments and news segment from three specific sources - Fox News, BreitbartNews and YouTube. The first batch (45,000 manually annotated) is ready for publication. We are currently in the final phase of manually labeling the remaining dataset using Amazon Mechanical Turk. BERT has been used widely in several downstream tasks. In this work, we present hBERT, where we modify certain layers of the pretrained BERT model with the new Hopfield Layer. hBert generalizes well across different distributions with the added advantage of a reduced model complexity. We are also releasing a JavaScript library and a Chrome Extension Application, to help developers make use of our...
Bi-directional Shape Correspondences (BSC): A Novel Technique for 2-d Shape Warping in Quadratic Time?
Uploads
Papers by Abdulrahman Ibraheem